WIM H. VAN BRAKEL*, ALISON M.

ANDERSON**,
STEPHEN G. WITHINGTON***, RICHARD P. CROFT

,
PETER G. NICHOLLS

, JAN H. RICHARDUS

&
W. CAIRNS S. SMITH

*KIT Leprosy Unit, Wibautstraat 137 J, 1097 DN Amsterdam,

Netherlands
**International Nepal FellowshipRELEASE, PO Box 28, Pokhara,
Nepal
***The Leprosy Mission Bangladesh, House 17A, Road 3,
Banani (Old) DOHS, Dhaka 1206, Bangladesh

56a St Peter's Road, Early, Reading RG6 1PH, UK

Department of Public Health, University of Aberdeen,

Polwarth Building, Foresterhill, Aberdeen, AB25 2ZD, UK

Department of Public Health, Erasmus MC,

University Medical Centre Rotterdam, PO Box 1738,
3000 DR Rotterdam, The Netherlands
Accepted for publication 19 August 2003
Summary This study was designed to investigate whether leprosy patients diag-
nosed with mild sensory impairment have a better prognosis when treated with
steroids than similarly impaired patients treated with placebo. A multi-centre,
randomized, double-blind, placebo-controlled trial was conducted in Nepal and
Bangladesh. Patients were eligible if they had a conrmed leprosy diagnosis, were
between 15 and 50 years old, had mild sensory impairment of the ulnar or posterior
tibial nerve of less than 6 months duration and did not require steroids for other
reasons. `Mild impairment' was dened as `impaired on the SemmesWeinstein
monolament test, but testing normal on the ballpen sensory test'. Subjects were
randomized to either prednisolone treatment starting at 40 mg per day, tapering over 4
months, or placebo. Nerve function was monitored monthly. Any patient who
deteriorated was taken out of the trial and was put on full-dose steroid treatment.
Outcome assessment was done at 4, 6, 9 and 12 months from the start of the treatment.
Outcome measures were the proportion of patients needing full-dose prednisolone
and the SemmesWeinstein sum scores. Each patient contributed only one nerve to
the analysis. Seventy-ve patients had nerves eligible for analysis, of whom 41 (55%)
Lepr Rev (2003) 74, 300310
The prognostic importance of detecting mild
sensory impairment in leprosy: a randomized
controlled trial (TRIPOD 2)
Correspondence: W. H. van Brakel (e-mail: w.v.brakel@kit.nl)
300 0305-7518/03/064053+11 $1.00 q Lepra
and 34 (45%) were allocated to the prednisolone and placebo arms, respectively. At 4
months, three patients in the prednisolone arm (7%) and six in the placebo arm (18%)
had an outcome event requiring full dose steroids. At 12 months, these proportions
had almost reversed, 11 (27%) and 6 (18%) in the treatment and placebo arms,
respectively. In the latter group, 75% had recovered spontaneously after 12 months.
Prednisolone treatment of sensory impairment of the ulnar and posterior tibial nerves
detectable with the monolament test, but not with the ballpen test, did not improve
the long-term outcome in terms of recovery of touch sensibility, not did it reduce the
risk of leprosy reactions or nerve function impairment beyond the initial 4-month
treatment phase. Two unexpected main ndings were the strong tendency of mild
sensory impairment to recover spontaneously and the fact that patients with mild
sensory impairment without any other signs or symptoms of reaction or nerve
function impairment are relatively rare.
Introduction
Testing of touch/pressure sensibility is a sensitive method for detecting nerve function
impairment in leprosy.
13
However, there is considerable debate about the best method of
sensory testing. Most of the world's leprosy patients are treated in peripheral health clinics in
resource-poor countries. Methods of clinical examination therefore need to be simple and the
instruments used readily available. Indentation of the skin is currently believed to be the most
quantiable way of measuring touch perception.
4,5
A simple technique of skin indentation that has found widespread acceptance in the
`eld' is the `ballpen test'.
6,7
The strength of the ballpen test is the universal availability of the
testing instrument: the ballpen. Usually ten sites are tested on each palm and sole. The result
is recorded as `felt' or `not felt' for each site tested. The main weaknesses of the test are its
lack of application force control and the fact that even very light application already exceeds
the human touch sensibility threshold by several multiples.
8
Despite this, the repeatability of
ballpen testing was moderate to good in the hands of trained staff.
9,10
The SemmesWeinstein monolaments provide an attractive alternative that is cheap,
and can provide a (semi-)quantitative result. Graded nylon monolaments are reported to be a
sensitive and repeatable method to detect mild sensory impairment (SI) in leprosy,
1,2,1114
and to monitor the treatment response of SI. The validity of the SemmesWeinstein
monolaments is well established. Johansson et al. showed that touch thresholds found
with monolaments correlate well with skin indentation thresholds measured with electronic
testing equipment.
4
In compression neuropathy, the monolaments correlated very well with
measurements of sensory bre conduction and electronic vibrometry.
15,16
The crucial question is whether the monolament test is `better' in detecting SI in leprosy
than the ballpen method. The underlying question is one of prognosis. `Do the patients in
whom neural impairment is detected at a mild stage with monolaments have a better
treatment prognosis than those whose impairment is detected with the ballpen test?' Since
there is insufcient information in the published literature, a trial was designed to answer this
question. The question was rephrased to, `Is there a signicant difference in treatment
outcome, in patients diagnosed with mild SI (monolament positive, but ballpen test
negative), between those who receive a standard steroid treatment regimen and those with
similar impairment who receive placebo treatment'.
TRIPOD 2 trial 301
The following hypotheses were formulated:
Patients with mild sensory impairment will get worse if left untreated, i.e. `monolament
sensory loss' will progress to `ballpen sensory loss' if left untreated.
Patients in the placebo group will become `ballpen test positive' signicantly more often
than those in the prednisolone group.
Materials and methods
DESIGN
The study was a multi-centre double blind, placebo controlled eld trial in six centres in
Nepal and Bangladesh. Subjects were randomized to receive prednisolone or placebo for a
duration of 16 weeks. The trial was part of a series of three trials (TRIPOD) which all used
very similar design and methods. The other two trials tested the use of prophylactic steroids to
prevent immunological reactions and peripheral neuropathy in leprosy and the steroid
treatment of neuropathy with a duration of more than 6 months before the start of treatment.
STUDY SUBJECTS AND SAMPLE SI ZE
All newly diagnosed MB patients attending one of the trial centres, who were not eligible for
the prophylactic steroids trial, were potentially eligible for the trial. They were included if
they had a conrmed MB leprosy diagnosis (positive skin smear or six or more skin lesions),
were between 15 and 50 years old, had sensory impairment of the ulnar or posterior tibial
nerve on the SemmesWeinstein test of less than 6 months duration, a normal ballpen test,
did not need steroids for any other reason, did not have a contraindication to steroids for any
reason and were living within a pre-dened geographical area so as to make follow-up by
home visit possible.
Based on previous experience, around 40% of patients with varying degrees of sensory
impairment regain `good' function after steroid treatment, when the impairment was
diagnosed with a `heavy' lament (,10 g for the hand, ,75 g for the foot). To detect an
improvement of 50% (that is change from 40% to 60%) with a power of 80% and with a
single sided 95% condence level, the trial needed approximately 80 subjects in each arm.
TREATMENT REGI MEN AND FOLLOW-UP
Subjects were randomized to either prednisolone treatment starting at 40 mg per day, tapering
over 4 months, or placebo. Nerve function was monitored monthly. Any patient who
deteriorated was taken out of the trial as a `bad outcome' and was put on full-dose steroid
treatment. Outcome assessment was done at 4, 6, 9 and 12 months from the start of the
treatment. Subjects developing an outcome event were taken out of the trial as `poor
outcome' and put on full-dose steroids.
EXAMINATION
The clinical assessment included a history, clinical examination for signs of reaction or
neuritis and a voluntary muscle test (05 grading). Outcome assessment (sensory testing) was
done using ve coloured graded monolaments: blue (200 mg), purple (2 g), red (4 g), orange
W. H. Van Brakel et al. 302
(10 g) and pink (300 g). The blue, purple, red and pink laments were used for the hand; the
purple, red, orange and pink for the foot. The test sites are shown in Figure 1. The ballpen test
was carried out on the same sites giving a light touch with the tip of the ballpen (just enough
to indent the skin very slightly). The patient was asked to indicate, with their eyes closed,
TRIPOD 2 trial 303
Figure 1. Monolament and ballpen test sites and hands and feet.
whether they felt the stimulus by pointing to where they felt the touch or by counting the
number of stimuli felt.
OUTCOME DEFI NITIONS
One point was given for every level that the monolament threshold was increased from
normal at each test site. The points were added for each nerve. Normal thresholds used were
200 mg for the hand and 2 g for the foot.
17,18
The SemmesWeinstein test was considered
positive if a patient scored 3 or more points for any nerve. If at a follow-up test a patient
scored the same score as at their baseline test, or 1 or 2 points less or more, then their
condition was considered `unchanged'. If the score had increased by 3 or more points, the
condition was diagnosed as `deteriorated'; if decreased by 3 or more points, it was diagnosed
as `improved'. If a patient's score improved by 3 or more points, and the total score for the
nerve was 2 or less then the patient's condition was called `recovered'.
The ballpen test was considered positive if 2 or more test sites did not feel the stimulus. A
bad outcome was dened as `worsening of sensory impairment so that the ballpen test
becomes positive'. A change in ballpen score was dened as a change of 2 or more test sites
feeling or not feeling the stimulus, compared to the previous result.
STATI STI CAL METHODS
Data entry and analysis was done using Epi Info software, version 6.04.
19
Outcome analysis
was based on a comparison of proportions (e.g. proportion of subjects recovered) or median
monolament scores between the treatment and placebo groups. The statistical signicance
of the difference between proportions was tested, where relevant, using a Z-test. For each
nerve entered into the study, the change in monolament score between the follow-up and
registration was calculated. The signicance of the difference in median sensory scores
between the two groups was tested with the KruskalWallis test.
20
Of patients with recent bilateral SI, only the most affected limb was included. If limbs
were equally affected, the right side was included. The results for the ulnar and posterior tibial
nerves were pooled in the analysis. If the patient had SI of both the ulnar and posterior tibial
nerve(s), they contributed up to two nerves to the study. However, these were never both
included in the same analysis.
Results
Eighty-four patients were enrolled in the trial. Nine were lost to follow-up, so the analysis
was based on data from 75 subjects (41 in the prednisolone group and 34 on placebo)
(Table 1). The equal distribution across the treatment and placebo group of mean age (372
versus 362), proportion of women (27% in each) and mean monolament score at diagnosis
(53 versus 52), shows that randomization was effective. Table 2 shows outcome events at
the different time points during follow-up. At the end of the rst 4 months, 2/41 (5%) patients
in the prednisolone group and 3/34 (9%) in the placebo group had become ballpen test
positive. The difference is not statistically signicant. By the 12-month follow-up, three more
patients in the (former) prednisolone group had become `ballpen positive', bringing the total
to 12%. In the placebo group, no further ballpen positivity events occurred. It is interesting
that only 2/8 of the ballpen positivity events occurred during steroid treatment.
W. H. Van Brakel et al. 304
Type 1 reactions were the next frequent group of outcome events, occurring in 4/41 (10%)
and 1/34 (3%) patients in the prednisolone and placebo groups, respectively (Table 2). None
of these reactions occurred during steroid treatment. It is noteworthy that all patients coming
out of the trial in the placebo arm did so within the rst 4 months, suggesting that the mild
sensory loss was possibly the rst symptom of a problem that would deteriorate. By 6 months,
the prednisolone arm had caught up and by 12 months, more people in the prednisolone arm
had developed an outcome event. The relative risk was not signicant at any point during
follow-up (Table 3).
Table 4 compares the difference in baseline monolament score and the score at the
time of various outcome assessments in the placebo and prednisolone groups. The `nal
outcome' shows data for all subjects from 12-month follow-ups, including people who
TRIPOD 2 trial 305
Table 2. Details of outcome events during follow-up among patients in the TRIPOD 2 trial
Prednisolone Placebo
Ballpen Ballpen
Month In trial positive T1R Others Lost In trial positive T1R Others Lost
0 41 34
4 35 2 1 3 24 3 1 2 4
6 31 1 2 1 24
9 29 1 1 24
12 26 1 1 1 24
Total 5 4 2 4 3 1 2 4
Table 1. Comparison of subject characteristics between those
receiving prednisolone and those on placebo
Prednisolone Placebo
(n 41) (n 34)
Mean age 372 362
% women 27% 27%
Mean monolament score at start 53 52
Table 3. Summary of outcome events and relative risk of having an outcome event at
various follow-up times during the TRIPOD 2 trial
Outcome events by treatment
Outcome events Relative risk
Month (cumulative) Prednisolone Placebo (all data n 75)
0 0 41 34
4 9 (12%) 3 (7%) 6 (18%) 241 (065893)
6 13 (17%) 7 (17%) 6 (18%) 103 (038278)
9 15 (20%) 9 (22%) 6 (18%) 080 (032203)
12 17 (23%) 11 (27%) 6 (18%) 066 (027159)
were taken out of the trial because of a poor outcome, where available. This best
represents the overall outcome of the result in both groups. Improvement in sensory
scores was very similar in both groups, except during the rst 4 months. During this
period, scores improved signicantly more in the prednisolone group, with none of the 41
subjects showing a worsening in monolament score. In the placebo group, worsening of
sensory scores was recorded in 12%. These differences had disappeared by the 6-month
follow-up.
W. H. Van Brakel et al. 306
Table 4. Comparison of the difference in baseline monolament score and the score at the time of various outcome
assessments in the placebo group (n 34) and prednisolone group (n 41) in the TRIPOD 2 trial (a negative sign
indicates improvement in score)
Difference in scores
Five categories
a
Two categories
a
Month Placebo difference Prednisolone difference Signicance Signicance RR
4 Mean 19 Mean 33 P001 P005 258 (119560)
Median 2 Median 3
Range 8 to ,5 Range 6 to ,0
6 Mean 26 Mean 28 P084 P034 099 (046210)
Median 3 Median 3
Range 8 to ,5 Range 7 to ,4
9 Mean 19 Mean 26 P022 P022 169 (086331)
Median 2 Median 3
Range 7 to ,5 Range 7 to ,4
12 Mean 25 Mean 26 P056 P040 134 (062292)
Median 2 Median 3
Range 9 to ,5 Range 7 to ,5
Final
b
Mean 30 Mean 27 P090 P041 106 (043261)
Median 3 Median 3
Range 9 to ,3 Range 7 to ,5
a
This score was categorized either in ve groups (worse, same, better, recovered (i.e. same as before the current
impairment) and normal (i.e. no longer an impairment) or in two groups (better (all improvement groups) and
`same/worse').
b
The `nal outcome' includes data from 12-month follow-ups for people out of the trial, and is therefore the
closest to the `steroid outcome' described in the background. In total 17 patients came out of the trial before the
12-month follow up. There were 13 such `post-trial-out' follow-ups.
Figure 2. Occurrence of outcome events during follow-up in the TRIPOD 2 trial.
The occurrence of all outcome events combined is shown in Figure 2. In the
prednisolone group, fewer events occurred during the treatment phase. However, 2
months after stopping steroids, the number of events has caught up with that in the placebo
group. At 12 months, more subjects in the prednisolone group have experienced an event
needing (additional) steroid treatment, although the difference is not statistically signicant
(P052).
Figure 3 compares the percentage of patients whose sensory function, as measured with
the monolament test, recovered or deteriorated between the prednisolone and placebo
groups. During the rst four months, a larger proportion of those on steroid treatment
recovered (71% versus 47%). This difference is close to signicance at the 5% level
(P0061). However, by 12 months the apparent disadvantage of the placebo group has
completely disappeared. In both groups close to 75% of patients have now recovered and only
a small number have worsened.
Figure 4 shows the progression of the mean MF scores during the different stages of the
trial. For patients who had two nerves registered in the trial, a separate analysis was done for
the rst and second nerve. It can be seen that the pattern of improvement in sensory scores is
TRIPOD 2 trial 307
Figure 3. Changes in sensory scores during steroid treatment (4 months) and after 12 months follow-up (n 75).
Figure 4. Mean monolament scores at various outcome time points during follow-up. A separate analysis was done
for subjects with more than one nerve involved in the trial.
very similar in both treatment and placebo categories. Improvement is slightly more gradual
in the placebo-treated nerves, but the end results are very similar.
Discussion
The purpose of this study goes back to the question, `What is the best sensory test to screen
for nerve function impairment under eld conditions?' The answer to this question depends
not only on scientic considerations, but also on many operational factors. From a scientic
point of view, `best' would refer to `best for the patient', i.e. improving the prognosis of any
neuropathy detected. The monolament test is believed to detect `early' or mild sensory
impairment, when the detection threshold is referenced to the normal threshold of sensa-
tion.
2,16,17,21
In contrast, the ballpen test may only detect sensory impairment at a more
advanced level, because the pressure exerted with a ballpen is variable and often much higher
than the normal threshold for touch sensibility. Actual data on the pressure exerted during
ballpen testing is difcult to obtain and we are not aware of any published reports on this.
Assuming that the SemmesWeinstein test would detect sensory impairment earlier than
the BP test, the question is, `would this early detection in fact improve prognosis?' In other
words, do people in whom sensory impairment was detected with the SemmesWeinstein
test and who were treated as soon as possible have a better prognosis than those in whom such
impairment is not treated until they become `ballpen positive'?
The results from the present trial would suggest that this is not the case. While patients in
the prednisolone-treated group improved slightly faster and did not show any deterioration
while on steroids, those in the placebo group showed spontaneous recovery which after 12
months matched that of the former (75% versus 73% in the steroid group). After 12 months,
there was no difference in mean or median monolament sensory scores between those
treated `early' and those treated only when developing `ballpen sensory impairment' (22
versus 16, respectively). In addition, the sensory function of some patients deteriorated after
the steroids had been stopped. In both groups, sensory impairment deteriorated in only a small
proportion of patients (7% versus 6%). Therefore, the current data do not indicate an
advantage in detecting sensory impairment before a carefully conducted ballpen test becomes
abnormal.
Whether more prolonged steroid treatment would have prevented the observed
deterioration, or indeed, the occurrence of the other outcome events observed during
follow-up, remains open to speculation. Many would nowadays consider a 12-week course
of prednisolone too short, particularly when treating MB patients. A recently conducted
prophylactic steroids trial (TRIPOD1) showed a similar pattern of results. Reactions and
nerve function impairment were prevented to a large degree during the steroid prophylaxis
period of 4 months, but the treatment group had an increased rate of events after stopping the
prophylaxis (Anderson et al., in preparation). By 12 months, the difference between the
prophylaxis and the placebo groups was no longer statistically signicant.
It should be noted that `ballpen positivity' did not always match deterioration as
measured with the monolament test (data not shown). Some patients had worsening
SemmesWeinstein thresholds, while the ballpen test remained negative; in others the
ballpen test became positive without deterioration in SemmesWeinstein scores. This is
likely to be due to the inherent variability in the application force of the ballpen. In
comparative studies, the test repeatability of monolaments was found to be better than
W. H. Van Brakel et al. 308
that of the ballpen test.
9,10
However, the results of the latter were still moderate to good,
suggesting that the test will give acceptable results if used carefully by trained staff.
The trial aimed to recruit four times as many patients as the eventual study size. An
interesting lesson from this trial was that patients with isolated mild sensory impairment were
difcult to nd. Most patients with sensory impairment also had other reasons for needing
steroids and therefore were not eligible to be included. These reasons included sensory
impairment of other nerves (e.g. the median nerve), motor impairment, severe type 1 or ENL
reaction or severe neuritis without NFI.
Some people would argue that the mild sensory impairment detected with the monola-
ment test is an artefact, caused by the subjective response of the person tested to a very small
stimulus, i.e. such diagnoses are false positives resulting from random error in the test. If this
were true, one would expect results similar to those observed in the current trial, namely no
difference in prognosis between the treatment and placebo group and spontaneous `recovery'
in the majority of subjects due to the phenomenon of `regression to the mean'.
20
However,
there are two arguments that plead against the hypothesis of mild sensory impairment being
false positive. The rst is the excellent repeatability that has been found in both experimental
and clinical studies with the monolaments.
9,10,13,14,22,23
Reliability coefcients were often
in excess of 090. This pleads against random error playing a major role in monolament
testing. The second argument comes from the observations made in this trial. While the
12-month results were almost equal in the treatment and placebo groups, there was a distinct
difference between the two during the time that subjects in the former group were taking
prednisolone. First, the sensory function in these subjects recovered faster (Figure 3). Second,
and perhaps most convincing, is the observation that the monolament sensory scores did not
deteriorate in any of the 41 subjects during steroid treatment, while deterioration was
observed in 4 out of 34 subjects in the placebo group (12%). The difference is statistically
signicant (P001). The above would indicate that the changes in sensibility measured with
the monolaments are genuine.
In conclusion:
A short (12-week) prednisolone course to treat sensory impairment of the ulnar and
posterior tibial nerves detectable with the monolament test, but not with the ballpen test,
did not improve the long-term outcome in terms of recovery of touch sensibility.
Mild sensory impairment has a strong tendency towards spontaneous recovery.
Patients with mild sensory impairment without any other signs or symptoms of reaction or
nerve function impairment are relatively rare.
Acknowledgements
TRIPOD was the work of the staff of the Biratnagar sub regional referral centre, the
Chittagong leprosy control project, the Danish-Bangladesh Leprosy Mission, Nilphamari, the
Dhaka leprosy control project, the Nepal Leprosy Trust Lalgadh and the Western Region
leprosy control project, Nepal.The successful completion of the trial is due to these people,
and to the patients themselves. This trial was sponsored by Lepra UK, TLMI, ALM, the
University of Aberdeen (UK) and the International Nepal Fellowship. The successful
completion of the trial is due to these people, and to the patients themselves. May the results
of this work bring glory to God.
TRIPOD 2 trial 309
References
1
Naafs B, Dagne T. Sensory testing:a sensitive method in the follow-up of nerve involvement. Int J Lepr, 1977; 45:
364348.
2
van Brakel WH, Shute J, Dixon JA, Arzet H. Evaluation of sensibility in leprosycomparison of various clinical
methods. Lepr Rev, 1994; 65: 106121.
3
Brown TR, Kovindha A, Wathanadilokkol U et al. Leprosy neuropathy: correlation of clinical and electro-
physiological tests. Ind J Lepr, 1996; 68: 114.
4
Johansson RS, Vallbo AB, Westling G. Thresholds of mechanosensitive afferents in the human hand as measured
with von Frey hairs. Brain Res, 1980; 184: 343351.
5
Peripheral Neuropathy Association. Quantitative sensory testing: a consensus report from the Peripheral
Neuropathy Association. Neurology, 1993; 43: 10501052.
6
WHO. A guide to leprosy control 2nd edition. World Health Organization,Geneva, Switzerland, 1988.
7
Watson JM. Disability control in a leprosy control programme. Lepr Rev, 1989; 60: 169177.
8
Bell-Krotoski JA, Buford JW. The force/time relationship of clinically used sensory testing instruments. J Hand
Ther, 1988; JanMar: 7685.
9
Lienhardt C, Currie H, Wheeler J. Inter-observer variability in the assessment of nerve function in leprosy patients
in Ethiopia. Int J Lepr, 1995; 63: 6276.
10
Anderson AM, Croft RP. Reliability of SemmesWeinstein monolament and ballpoint sensory testing, and
voluntary muscle testing in Bangladesh. Lepr Rev, 1999; 70: 305313.
11
Hammond CJ, Klenerman P. Protective sensation in the foot in leprosy. Lepr Rev, 1988; 59: 347354.
12
Brown TR, Kovindha A, Wathanadilokkol U et al. Clinical assessment of early leprous neuropathy with
electrophysiologic correlation. Int J Lepr, 1993; 61: 148A.
13
van Brakel WH, Khawas IB, Gurung KS et al. Intra- and inter-tester reliability of sensibility testing in leprosy. Int
J Lepr, 1996; 64: 287298.
14
Birke JA, Brandsma JW, Schreuders TA, Piefer A. Sensory testing with monolaments in Hansen's disease and
normal control subjects. Int J Lepr, 2000; 68: 291298.
15
Szabo RM, Gelberman RH, Williamson RV et al. Vibratory sensory testing in acute peripheral nerve compression.
J Hand Surg, 1984; 9A: 104109.
16
Gelberman RH, Szabo RM, Williamson RV, Dimick MP. Sensibility testing in peripheral-nerve compression
syndromes. An experimental study in humans. J Bone Jt Surg [Am.], 1983; 65: 632638.
17
Kets CM, van Leerdam ME, van Brakel WH et al. Reference values for touch sensibility thresholds in healthy
Nepalese volunteers. Lepr Rev, 1996; 67: 2838.
18
Anderson AM, van Brakel WH. Age specic normal thresholds for sensibility testing with monolaments in a
Nepali population. Int J Lepr, 1998; 66: 69A.
19
Dean AG, Dean JA, Dicker RC. Epi Info: a word processing, database,and statistics program for public health.
(6.03). 1996. Centres for Disease Control & Prevention, Atlanta.
20
Altman DG. Practical statistics for medical research. Chapman & Hall, London, 1991.
21
Bell-Krotoski JA, Ewing FE, Figarola JH, Hiltz D. Threshold detection and SemmesWeinstein monolaments.
J Hand Ther, 1995; AprJune: 155162.
22
Bell-Krotoski JA, Tomancik E. The repeatability of testing with SemmesWeinstein monolaments. J Hand
Surg, 1987; 12A: 155161.
23
Griener JS, Muro LB, Jones SV. Threshold of sensation; interrater reliability and establishment of normal using
the SemmesWeinstein monolaments. J Hand Ther 1989.
W. H. Van Brakel et al. 310